Microischemia, oxidative stress, and glutamate excitotoxicity are likely initiators of age-related neurodegenerative diseases. Many senile dementias are characterized by a loss of synapses (up to 50 percent loss compared to age related controls) within the hippocampus and cortex, whereas neuronal loss is far less. The mechanism by which some neuronal processes selectively degenerate while associated cell bodies and other processes remain intact is largely unexplained. Treatment of cultured hippocampal and cortical neurons with common mediators of neurodegeneration leads to the formation of axonal and dendritic inclusion bodies in the form of actin depolymerizing factor (ADF)/cofilin-cortactin-actin rods. ADF/cofilins are essential proteins that regulate the turnover of actin in vivo. Through their regulation by phosphorylation on a single serine residue by LIM kinase, ADF/cofilins are a common target of signaling pathways for actin cytoskeletal reorganization via the rho family of GTPases. Cortactin normally associates with the cortical actin cytoskeleton and is regulated by non-receptor tyr kinases. It has domains for binding Arp2/3 complex and Factin and may cross-link filaments in the rods. Treatments that induce rods cause ADF/cofilin dephosphorylation and at least a transient inactivation of mitochondria. Rod-like inclusions containing ADF are found in Alzheimer brain: half of the inclusions occur nearby amyloid plaques, but >98 percent of plaques have neurites nearby that contain inclusions. Normal human brains do not contain similar ADF-inclusions. Growth cones disappear from neurites containing persistent rods but normal growth cones are found on other processes extending from the same soma, confirming that persistent rod formation is accompanied by degeneration of the neurite distal to the inclusion. Rods may provide a mechanism linking mitochondrial dysfunction to the selective pruning of synaptic terminals. Here we propose to determine the role of rods in synapse elimination. Our specific aims are to test the following interrelated hypotheses: Formation of a persistent rod in a neurite process inhibits transport to the distal regions of that neurite. Synaptic function in a process distal to a rod will be impaired. Amyloidpeptides and plaque will induce rods and therefore rods will contribute to the pathology in mice expressing the mutant amyloid precursor protein (APP v717F). Rod formation requires cortactin dephosphorylation and cortactin domains that bind Arp2/3 complex and F-actin.